Back to skill

Security audit

MLOps Observability

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small MLOps observability guide with an MLflow example; its behavior is disclosed and fits its stated purpose.

Before installing, note that the included MLflow helper can log model artifacts, run parameters, dataset-derived metadata, and git commit identifiers. Review what your training config contains before using it with sensitive data, and clarify the unexplained '-cn' naming if regional scope matters to your workflow.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Natural-Language Policy Violations

Medium
Confidence
84% confidence
Finding
The manifest-style header names the skill `mlops-observability-cn`, which implies a locale or regional constraint. The file does not explain whether this China-specific designation is optional, user-selected, or justified by a region-specific compliance need, which can violate the language/locale policy requirement.

Natural-Language Policy Violations

Low
Confidence
82% confidence
Finding
The package name includes the locale marker "-cn", which indicates a China-specific variant. In this manifest file there is no accompanying explanation, opt-in, or documented regional justification, so the natural-language metadata suggests a locale constraint without clear user choice.